Scores.wiki · Research

272 NFL betting angles, tested

Every angle gamblers repeat, run against all 7,276 games from 1999 to 2025. What each angle is, how you would play it, and what it actually returned — split before and after 2018.

7,276 games · 14,552 team-games · nflverse · break-even at −110 = 52.4%

272Angles tested

4Were look-ahead leaks

0Beat the moneyline price

1Still standing

How to read this

The bar is 52.4%, not 50%

At −110 you risk 110 to win 100, so a bet must hit 52.4% just to break even. A 51% angle loses money. ROI below is per unit staked at a flat −110; pushes count as zero. 2018 is the split because PASPA fell that May and the money changed — "2018 onward" includes 2018.

Dead no edge, either era Watch positive both eras, still unproven Leak only worked using information you would not have had

The angles

Every angle, and what it returned

Nothing below clears the bar in both eras with a sample worth trusting. The two marked Watch are positive but not statistically separable from break-even.

Rest & the bye week

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Home favourite off a byeBack the rested home favourite	219	50.7%	-3.2%	78	41.0%	-20.9%	Dead
Any team off a byeBack the rested team	609	51.7%	-1.2%	249	47.4%	-9.3%	Dead
Facing a rested opponentBack the team that did not have the bye	609	48.3%	-7.7%	249	52.6%	+0.4%	Dead
Short week (Thursday)Back the team on 4 days' rest	316	50.0%	-4.5%	251	49.8%	-4.8%	Dead
Big rest advantage (7+ days)Back the better-rested side	468	52.4%	-0.1%	165	45.5%	-13.1%	Dead
Mini-bye (10 days' rest)Back the team off the mini-bye	369	50.4%	-3.7%	292	51.0%	-2.5%	Dead
Third straight road gameBack the road team on a long trip	76	46.1%	-12.1%	27	48.1%	-7.8%	Dead
Game before the byeBack the team playing into its bye	585	51.8%	-1.1%	253	56.9%	+8.5%	Dead

Travel & body clock

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
West-coast team, early East kickoffBack the travelling Pacific team (body clock 10am or earlier)	184	47.3%	-9.4%	109	56.9%	+8.4%	Dead
Crossing three zones eastwardBack the eastbound traveller	264	48.9%	-6.5%	153	59.5%	+13.1%	Dead
Crossing three zones westwardBack the westbound traveller	272	50.7%	-3.1%	144	55.6%	+5.9%	Dead
Trips over 2,000 milesBack the long-haul traveller	468	50.2%	-4.0%	212	54.7%	+4.4%	Dead
Visiting Denver (altitude)Back the visitor at Denver	106	55.7%	+6.3%	48	47.9%	-8.5%	Dead
Week after a London gameBack the team returning from London	13	92.3%	+70.8%	46	39.1%	-25.3%	Dead
Dome team playing outdoorsBack the dome team outdoors	875	49.8%	-4.8%	438	53.7%	+2.3%	Dead

Weather & surface

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Underdogs in 20mph+ windBack the underdog outdoors in heavy wind	143	53.8%	+2.8%	31	51.6%	-1.5%	Dead
Underdogs in 10–14mph windBack the underdog outdoors in moderate wind	824	48.8%	-6.7%	293	55.6%	+6.1%	Dead
The wind-under criteria, on the sideBack the underdog in the exact cell that produces the totals edge	482	51.5%	-1.7%	175	46.9%	-10.5%	Dead
Heavy favourites in 15mph+ windBack the big favourite in wind	150	50.0%	-4.4%	47	57.4%	+9.5%	Dead
Warm-weather team in the freezing coldBack the cold-climate host	80	57.5%	+9.7%	24	58.3%	+10.9%	Dead
Grass team forced onto turfBack the turf host	1,189	51.9%	-0.9%	527	50.7%	-3.2%	Dead

Form, streaks & revenge

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Fade a team on a 3+ ATS streakBack the opponent of the hot team	1,201	50.4%	-3.7%	541	49.2%	-6.0%	Dead
Back a team on a 3+ ATS losing runBack the cold team	1,157	51.7%	-1.3%	516	50.4%	-3.7%	Dead
Revenge gameBack the team that lost the first meeting	935	52.3%	-0.2%	371	51.5%	-1.7%	Dead
Letdown after a blowout winFade the team coming off a blowout win	1,020	50.7%	-3.1%	422	52.4%	-0.0%	Dead
Bounce-back after a blowout lossBack the team coming off a blowout loss	1,563	50.4%	-3.8%	671	50.7%	-3.2%	Dead
Buy low after a blowout loss + line moveBack a team whose line moved 3+ against it after a 14+ point loss	433	55.4%	+5.6%	194	54.1%	+3.3%	Watch
Overtime hangoverFade the team coming off an overtime game	546	48.7%	-6.8%	216	50.5%	-3.6%	Dead

Public money, teams & coaches

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Public teams as favouritesBack DAL/GB/NE/PIT/SF/KC/PHI when favoured	1,363	51.4%	-1.9%	654	49.5%	-5.3%	Dead
Fade public teams as favouritesBack the opponent of a public favourite	1,363	48.6%	-7.0%	654	50.5%	-3.6%	Dead
Public teams, home, primetimeBack the marquee brand at home under lights	202	54.0%	+3.0%	132	49.2%	-5.9%	Dead
Low-profile teams as underdogsBack the unfashionable underdog	1,475	50.1%	-4.2%	655	52.7%	+0.5%	Dead

Primetime & scheduling

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Monday night underdogsBack the underdog on MNF	312	49.0%	-6.2%	149	51.7%	-1.3%	Dead
Sunday night underdogsBack the underdog on SNF	493	50.1%	-4.2%	142	52.8%	+0.8%	Dead
Thursday night underdogsBack the underdog on TNF	180	41.7%	-19.9%	138	52.2%	-0.4%	Dead
Primetime home underdogsBack the home underdog under lights	271	45.0%	-13.6%	180	52.8%	+0.7%	Dead
Underdogs in weeks 1–4Back the underdog early in the season	1,119	52.5%	+0.1%	498	54.4%	+3.8%	Watch
Underdogs in weeks 15–18Back the underdog late in the season	878	51.1%	-2.3%	447	50.3%	-3.8%	Dead

Divisional games

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Divisional road underdogsBack the division rival on the road	1,207	53.2%	+1.5%	447	54.6%	+4.1%	Watch
Divisional underdogs (all)Back the division underdog	1,857	51.6%	-1.4%	754	53.1%	+1.3%	Dead
Divisional favouritesBack the division favourite	1,857	48.4%	-7.4%	754	46.9%	-10.2%	Dead

Price structure

Angle & how it is played	Pre-2018			2018 onward			Verdict
Angle & how it is played	n	Win%	ROI	n	Win%	ROI	Verdict
Home teams generallyBack the home side	4,703	49.1%	-6.1%	2,075	48.7%	-6.9%	Dead
Road underdogsBack the road underdog	3,107	51.2%	-2.1%	1,250	52.8%	+0.8%	Dead
Double-digit favouritesBack the favourite laying 10+	577	47.0%	-10.1%	268	49.6%	-5.1%	Dead
Home underdogsBack the home underdog	1,566	49.7%	-5.0%	825	50.9%	-2.8%	Dead
Big road underdogs (moneyline)Back the 7+ point road dog to win outright	711	16.9%	-19.6%	428	19.2%	-12.5%	Dead

The one that survives

The Wong teaser

What it is: move a favourite from −7.5/−8.5 down through both 7 and 3, or an underdog from +1.5/+2.5 up through both. Those are the two most common margins in football — games land on exactly 3 in 15.0% of cases and on 7 in 9.1%, and nothing else tops 6%.
How it is played: two legs, six points each, on one ticket. Both must win.

Two-team 6-point teaser	Pre-2018 n=795 · 74.1%	2018 onward n=515 · 74.6%	All n=1,310 · 74.3%
Priced −110	+4.8%	+6.1%	+5.3%
Priced −120	+0.6%	+1.9%	+1.1%
Priced −130	−2.9%	−1.6%	−2.4%

Real mechanism, stable, and the one angle that improved after 2018. But it is not proven: the confidence interval on the leg rate still includes break-even (p = 0.065), and the whole thing lives in the price. At −110 it is worth having; at −130 — the price many books moved to precisely because this play is well known — it loses. Check what your book charges before anything else.

Legs from the same week are correlated (ICC = 0.072), so pair legs from different weeks where you can.

Disqualified

The four that looked best were cheating

Each backtested well and was era-stable. Each used information that did not exist when the bet would have been placed. Rebuilt using only pre-game data, all four collapse.

Angle	Why it is invalid	As tested	Rebuilt honestly
Fade a favourite who is a dog next week	Next week's line is set after this week's game	57.4%	50.3%
Sandwich trap	Same — conditions on next week's line	59.4%	51.9%
Back one-quarterback teams	Counting a full season's starters needs the future	53.0%	50.5%
Fade quarterback-churn teams	Same full-season count	54.9%	51.2%

Caveats

What this cannot tell you

Most "Dead" verdicts mean "not found", not "not there". Detecting a genuinely bettable 2-point edge needs about 4,900 bets. Only 8 of the 272 tests had that much data; the median had 645. A 6–8 point edge could hide in most of these rows.
2018 changed nothing measurable. Of 240 angles with data on both sides, 13 shifted at p < 0.05 — against ~12 expected by chance.
The moneyline is efficient. Zero of 43 moneyline angles beat the market's own de-vigged price. Teams do not win outright more often than their price implies.
Closing lines only. No opening lines or line movement, so "market overreaction" ideas are tested weakly. No injury data, and no alternate-line prices.
No team and no coach beats the number by more than chance (p = 0.56 and 0.28).

Built on nflverse schedules via a team-game panel: 14,552 rows, 1999–2025, with travel, rest, form and market features. Form fields verified leak-free; all 272 tests including nulls retained so the multiple-testing correction stays valid. · scores.wiki